Woof wrote: > On Mon, 29 Sep 2008 15:24:06 -0400, Raymond Dans > <[EMAIL PROTECTED]> wrote: > > > After examining the code for these two areas, I cannot > figure out why > > this would still be the case after the fix for "XECS-1583 - > Imdb Table > > Replication sometimes times out with large table". With the fix > > (which is used by SipClient) > ... > > Does anyone know if XECS-1599 is still a valid issue or can > I close it > > off. > > I believe XECS-1599 was created for OTHER places where poll() > is used. It was opened as a "we found this bug in one place, > let's track down all the other places and make sure we do the > right thing" entry. > The problem wasn't in how we were using poll() but in how we reacted when we were told there was data available. Prior to fixing XECS-1583, we weren't properly emptying the socket buffer of all the data so when we went back to request a poll() it would tell us there was no data even though there was. Now that we completely empty the buffer, if poll() tells us that there is no "new" data then we can be sure that there isn't. The method used by the SipClient for reading from the buffer after poll() had indicated data is the same as was fixed in XECS-1583.
Raymond _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
